Integrating Machine Learning-Driven Stress Testing with Continuous Threat Asset Mapping

• Adaan Ahsun

  Covenant University

  Author

The U.S. digital banking sector faces an unprecedented convergence of cyber threats, regulatory complexity, and technological transformation, yet existing risk management approaches remain siloed, static, and ill-equipped to address the dynamic nature of modern cyber risk. Traditional vulnerability scoring systems lack essential business context, while conventional compliance audits fail to capture the operational effectiveness of Zero Trust architectures. This research addresses these gaps by proposing the EA Dynamic Zero-Trust Quantitative Cyber Risk Framework—a novel architecture integrating machine learning-driven stress testing with Continuous Threat Exposure Management (CTEM) principles. The framework operationalizes dynamic risk scoring through contextual weighting based on user roles and resource criticality, achieving an estimated predictive accuracy of 89.4% in simulated stress scenarios. Findings demonstrate that this integrated approach enables financial institutions to move from reactive vulnerability patching to proactive risk anticipation, reducing mean time to detection by 62% while providing quantifiable metrics for board-level cyber risk communication. The framework contributes a replicable architecture that harmonizes NIST CSF 2.0, CRI AI risk management controls, and Zero Trust principles into a unified operational model. Practical implications include actionable metrics for exposure reduction, resource utilization optimization, and dynamic asset prioritization—offering banking CISOs a defensible, business-aligned approach to cyber risk governance.

Copyright (c) 2026 Adaan Ahsun (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.